src/cpu/x86/vm/stubGenerator_x86_64.cpp

*** 142,153 ****
    //    72(rbp): thread                                 Thread*
    //
    //     [ return_from_Java     ] <--- rsp
    //     [ argument word n      ]
    //      ...
!   //  -8 [ argument word 1      ]
!   //  -7 [ saved r15            ] <--- rsp_after_call
    //  -6 [ saved r14            ]
    //  -5 [ saved r13            ]
    //  -4 [ saved r12            ]
    //  -3 [ saved rdi            ]
    //  -2 [ saved rsi            ]
--- 142,156 ----
    //    72(rbp): thread                                 Thread*
    //
    //     [ return_from_Java     ] <--- rsp
    //     [ argument word n      ]
    //      ...
!   // -28 [ argument word 1      ]
!   // -27 [ saved xmm15          ] <--- rsp_after_call 
!   //     [ saved xmm7-xmm14     ] 
!   //  -9 [ saved xmm6           ] (each xmm register takes 2 slots)
!   //  -7 [ saved r15            ] 
    //  -6 [ saved r14            ]
    //  -5 [ saved r13            ]
    //  -4 [ saved r12            ]
    //  -3 [ saved rdi            ]
    //  -2 [ saved rsi            ]
*** 167,178 ****
    //    We spill c_rarg0-c_rarg3 to this space.
  
    // Call stub stack layout word offsets from rbp
    enum call_stub_layout {
  #ifdef _WIN64
!     rsp_after_call_off = -7,
!     r15_off            = rsp_after_call_off,
      r14_off            = -6,
      r13_off            = -5,
      r12_off            = -4,
      rdi_off            = -3,
      rsi_off            = -2,
--- 170,184 ----
    //    We spill c_rarg0-c_rarg3 to this space.
  
    // Call stub stack layout word offsets from rbp
    enum call_stub_layout {
  #ifdef _WIN64
!     xmm_save_first     = 6,  // save from xmm6
!     xmm_save_last      = 15, // to xmm15
!     xmm_save_base      = -9,
!     rsp_after_call_off = xmm_save_base - 2 * (xmm_save_last - xmm_save_first), // -27
!     r15_off            = -7,
      r14_off            = -6,
      r13_off            = -5,
      r12_off            = -4,
      rdi_off            = -3,
      rsi_off            = -2,
*** 206,215 ****
--- 212,228 ----
      parameter_size_off =  2,
      thread_off         =  3
  #endif
    };
  
+ #ifdef _WIN64
+   Address xmm_save(int reg) {
+     assert(reg >= xmm_save_first && reg <= xmm_save_last, "XMM register number out of range");
+     return Address(rbp, (xmm_save_base - (reg - xmm_save_first) * 2) * wordSize);
+   }
+ #endif
+ 
    address generate_call_stub(address& return_address) {
      assert((int)frame::entry_frame_after_call_words == -(int)rsp_after_call_off + 1 &&
             (int)frame::entry_frame_call_wrapper_offset == (int)call_wrapper_off,
             "adjust this code");
      StubCodeMark mark(this, "StubRoutines", "call_stub");
*** 254,265 ****
      __ movptr(rbx_save, rbx);
      __ movptr(r12_save, r12);
      __ movptr(r13_save, r13);
      __ movptr(r14_save, r14);
      __ movptr(r15_save, r15);
- 
  #ifdef _WIN64
      const Address rdi_save(rbp, rdi_off * wordSize);
      const Address rsi_save(rbp, rsi_off * wordSize);
  
      __ movptr(rsi_save, rsi);
      __ movptr(rdi_save, rdi);
--- 267,281 ----
      __ movptr(rbx_save, rbx);
      __ movptr(r12_save, r12);
      __ movptr(r13_save, r13);
      __ movptr(r14_save, r14);
      __ movptr(r15_save, r15);
  #ifdef _WIN64
+     for (int i = 6; i <= 15; i++) {
+       __ movdqu(xmm_save(i), as_XMMRegister(i));
+     }
+ 
      const Address rdi_save(rbp, rdi_off * wordSize);
      const Address rsi_save(rbp, rsi_off * wordSize);
  
      __ movptr(rsi_save, rsi);
      __ movptr(rdi_save, rdi);
*** 358,367 ****
--- 374,388 ----
        __ bind(L);
      }
  #endif
  
      // restore regs belonging to calling function
+ #ifdef _WIN64
+     for (int i = 15; i >= 6; i--) {
+       __ movdqu(as_XMMRegister(i), xmm_save(i));
+     }
+ #endif
      __ movptr(r15, r15_save);
      __ movptr(r14, r14_save);
      __ movptr(r13, r13_save);
      __ movptr(r12, r12_save);
      __ movptr(rbx, rbx_save);